4 systems thinkinh in agri

8/20/2019 4 Systems Thinkinh in Agri

1/27

1

4 System thinking in agriculture: an overview systemmostfinal2

J.B. Schiere1, R. Groenland

2, A. Vlug

3 and H. Van Keulen

4

Chapter 4 in: System Thinking in Agriculture an overview. Chapter 4 in Emerging Challenges

for farming systems - lessons from Australian and Dutch agriculture; edited by Ken Rickert.Rural Industries Research and Development Corporation. P.O.Box 4776, Kingston Act 2604.

A sufficiently vast and considerable intellect, knowing the complete laws of microscopic physics and

the physical state of the universe at any moment, could calculate the state of the universe at all

subsequent times; all prior ones, too, if the laws are reversible. (Laplace, Essai philosophique sur les

probabilités, 1814)

Step by step, there lies waiting, quite perfect, what is needed for that step (Indian wisdom, quoted by

Auerbach, 1999);

A reference to Aristotle and Leibniz has long ceased to be required in serious books. But […] severalbasic ideas of fractals might be viewed as mathematical and scientific implementations of loose but

potent notions that date back to Aristotle and Leibniz, permeate our culture, and affect even those

who think they are not subject to philosophical influences. (Mandelbrot, 2000).

The road emerges as we are walking (a line from a Spanish poem by Antonio Machado)

„Maturity is the ability to hold on to different world views, preferably conflicting, at the same time.‟

(C. West Churchman quoted by Richard Bawden)

„Newton separated light into parts [different colours] while Goethe was int erested to see what

happened when parts [different colours] were joined together‟ (Brian Goodwin, pers. comm., 2000).

Abstract

Agricultural developments in the 20th century are characterised by impressive yield increases. In spite

of its successes, however , agriculture‟s image has also become marred by issues such as unequal

distribution of food and income, large-scale social change and concerns about pollution and/or

exhaustion of natural resources. Can agriculture continue to feed the world, and if so, in which way? All

these issues relate to similar developments and concerns in other sectors of society and in our opinion

system thinking provides concepts and tools for the analysis and design of such developments. System

thinking is considered here as an informal kind of system theory that can also use emerging and as yet

immature concepts to explore the behaviour of the world in which we live. System thinking may takemany different forms and can help to explore questions regarding the future of farming, to the extent

that agricultural systems represent a special case of systems in general. Much can also be learned by

combining insights from agriculture with those from other disciplines, such as mathematics, physics and

psychology. Such combinations should even allow for the bridging of traditional – but man-made –

divisions. Therefore, this chapter outlines different forms of system thinking. It discusses aspects of

reductionist and holistic approaches, static and dynamic thought, as well as the so-called control versus

1 Hans Schiere, La Ventana ASA&D, Bovenweg 30, 6721 HX Bennekom, the Netherlands; www.laventana.nl2 Rob Groenland, Meteo Consult bv, Agro Business Park 99/101 6708 PV Wageningen, the Netherlands;

[email protected] Albert Vlug , Erasmus University, Medical Informatics, P.O. Box 1738, 3000 DR Rotterdam, the

Netherlands; [email protected] Herman Van Keulen ,Plant Research International, P.O. Box 16, 6700 AA Wageningen, the Netherlands:

[email protected]


2/27

2

participation paradigm. We think that these extremes span „ reality‟ , and we therefore opt for a

combination of approaches rather than for one or the other. This is the central issue of this chapter, i.e.

the need to integrate insights from different disciplines, from thinking focussed on matter to thinking

focussed on mind, from arts to sciences. By doing so, it aims to show that concepts from psychology,

ecology and physics, such as perception, coping strategies, predator-prey relations and thermodynamic

theory, can clarify discussions about agricultural development. It also aims to show that man is part ofnature, that „ he‟ is a participant that has to make choices rather than a neutral observer who can control

nature for a single-minded anthropocentric goal. The last part of this chapter is therefore dedicated to

the discussion of generalised forms of system behaviour and their application in daily life, based on

thermodynamic theory.

Introduction

Agricultural development of the past century tends to be characterised by impressive production

increases through increased use of land, fossil resources, labour, and technological innovations or

farmers‟ ingenuity (Chapters 1 and 2). The achievements, however, are marred by negative trade -offs

that seem to be intimately linked to so-called progress. Examples of emerging concerns in this respect

are unequal distribution of food and income at local, regional and global levels, declining bio-diversity,

dwindling water reserves, and side-effects from using biotechnology etc. Some authors maintain a rosy

or cornucopian view of the options for agriculture, others are more conservationist, while an

intermediate group takes a balanced but necessarily ambiguous view. The divergence of opinion is

confusing in itself, but matters are made even more complex due to the fact that changes in agriculture

cannot be understood in isolation, they are inextricably linked to other developments in society.

System theory provides concepts and tools to better understand complex developments in agriculture

and society, because farming systems are just one type of system in general. The terms „system theory‟

and „system thinking‟ both refer to an activity that is as old as mankind and that knows many traditions.

We mostly use the term „ thinking‟ because it permits the use of valuable ideas that are not (yet)

formalised. Understanding the different traditions in system thinking can clarify the origin of present practices and policies in agriculture; it can even have therapeutic value in showing the roots of „ hang-

ups‟ in modern thought, and it can help to establish choices for the future. This chapter, therefore,

outlines a few major traditions in western system thinking and ventures into their significance for the

21st century. In doing so we address three major issues:

Uncertainty and changing conditions. Three key concepts in this respect are context, relations

and non-linearity. Context is a general term for the management environment outlined in Chapter 1.

It implies that what is „ good‟ or true in one place may be „ bad‟ or untrue elsewhere. For example,

fossil fuels or agro-chemicals may be useful in one place but counterproductive in other systems.

The existence of relations implies that an action in one place may trigger unexpected effects

elsewhere. For example, the introduction of pesticides can result in trade-offs that cause consumers

to switch to other products. Non-linearity refers to the phenomenon that trends cannot beextrapolated linearly, and that too little as well as too much of any given intervention or activity

may not be good (Odum, 1975).

The link between matter and mind. Bridging traditional divides, such as exist between thinking in

terms of matter and mind , is in our view essential for a future that is worth living. Much thinking

about agriculture focuses on biophysical entities (matter) that can be quantified, thus leaving

qualitative, intangible and psychological (mind) effects exclusively to considerations of society and

politics. At the same time, many socially oriented disciplines have difficulty in understanding and/or

accepting the importance of physical laws. In this chapter we continue to use the distinction between

matter and mind to make sure that both aspects are considered; not to imply that they are separable

in some strict sense.


3/27

3

Enabling discussion. Concepts pertaining to system behaviour provide a basis for discussing the

future of agriculture, exploring uncertainty and variation in systems, linking cornucopian and

conservationist thought, and for reassessing basic notions in our thinking.

This chapter first reviews ancient traditions in system thinking. It subsequently distinguishes modern

schools of system thought, here categorised as hard-, soft- and complex system thinking, respectivelyHSM, SSM and CSM. (the „ M‟ in this acronym stands for „ methodology‟). It also makes a distinction

between approaches, thinking, and methodology, because thinking about approaches necessarily

precedes choice of method. Brief reference is made to farming systems research across the globe. We

discuss the relation between complexity, inevitable trade-offs and the strange mixture of uncertainty and

repetition of form. Finally, a discussion of system dynamics contrasts rather static thinking about

farming systems with evolutionary approaches. It thus connects major themes of modern and ancient

system thinking by relating several forms of system behaviour, like co-evolution, lock-in and predator-

prey relations to aspects of agricultural development. Whilst a few of these behaviours are known in one

way or another, the section on ancient system thinking shows that new concepts tend to be initially

accepted for their instrumental value, and only later lead to a paradigmatic shift in thinking if at all.

Useful background reading is found in Bertalanffy (1968), Stakman et al. (1967), Prigogine & Stengers

(1985), Gleick (1987), Cohen & Stewart (1994), Klir (1991), Ison & Russell (1999), Conway (1987)

Capra (1997), Röling (1996), Checkland (1999), Jackson (2000) and Collinson (2000).

Ancient system thinking and uncertainty

Western system thinking can trace its origins to the days before the ancient Greeks and today still tends

to repeat arguments resembling those of millennia ago. For example, a major rift in modern thought

occurs between traditions that think in terms of steady states (equilibrium, ceteris paribus), clockwork

notions and „ hard‟ facts on the one hand, and traditions that assume „ fluidity‟ and uncertainty (non-

equilibrium, ceteris imparibus) on the other. The first tradition sets fixed targets and tends to aim for

„ final‟ solutions. In this school of thought, an irrigation scheme, a new cultivar or biotechnology will,

once and for all, solve the world food problem: „the war against hunger must be won‟ (Stakman et al.,1967). The latter tradition thinks in fluid terms of change, combining several perceptions of reality, and

it accepts change by adaptive management. This approach reflects the notion of the ancient Greek

philosopher Heraclitus who said „nothing is permanent‟ . This view contrasted with that of the

Pythagoreans, who attempted to perceive everything in terms of geometry. Much to their dismay the

latter discovered that the value of („ pi‟ ) could not be calculated numerically, a shock that resembles

the difficulty of today's paradigm-shift: away from clockwork-certainty towards uncertainty; from

control to participation5. Some centuries later, the early church „ controlled‟ uncertainty for at least a

millennium by proposing dogmas as a form of „ unquestioned authority‟ . This approach was challenged

during the Renaissance by people who valued empirical observation over church authority, a challenge

reminiscent of the one today by postmodernists to traditional science (Funtowitz and Ravets, 1994). It

was a time when Galileo and Copernicus proposed a heliocentric model of the solar system, an ideaalready expressed by early Greeks such as Aristarchus. It allowed for the use of much simpler methods

to understand planetary behaviour. But at that time this was a notion that ran counter to established

dogma. Furthermore, Kepler‟s suggestion to use ellipses challenged the religious „ certainty‟ that planets

represented some kind of deity, which by virtue of this nature should follow circles as indicative of their

perfect form. It was one thing to „ instrumentally‟ use these formulas to more easily explain the motion

of the planets, it was something else altogether to suggest that they implied a paradigm shift, i.e. that

they should affect our thinking about the nature of the universe.

A major contribution of Descartes to the intellectual discourse concerning „ reality‟ and uncertainty was

to revive ancient notions of reason as a complement to the inadequacy of observation alone. As a father

5 The terms “control” or “participation” is taken from Brian Goodwin (pers. comm., 1999)


4/27

4

of modern reductionism he also rekindled a hard duality: the unequivocal separation of matter (res

extans) and mind (res cogitans). This distinction helped to cope with uncertainty, thus providing a

useful conceptual tool in a time of much mysticism and little critical observation. Unfortunately, this

separation of mind and matter has almost become an article of faith that is hard to shed. Today there

may be too much emphasis on matter and too little on mind in thought on agriculture, although some

biophysical thinkers tend to attribute to their socio-economic counterparts an excessive preoccupationwith mind.

Descartes was followed by Newton, who developed formulas that explained what Leibniz described as

the „ clockwork ‟ universe, in which everything is predictable i.e. where „ certainty‟ rules, where „ control‟

is possible, and where one knows „ what happens in one place if one presses a lever in another place‟ .

However, Newton was more proud to have discovered similarity between the microcosm of the falling

apple and the macrocosms of planets. He was not primarily interested in the „ clockwork ‟ aspect of the

universe. Perhaps his interest in this respect was more incidental, and more in the repetition of form

sense that is called fractal behaviour and discussed later in this chapter (B ox 4.1.). Newton‟s writings

became the prototype for scientific reasoning and his axioms nourished Laplace‟s fantasy about the

omniscient scientist: a „daemon‟ knowing the laws of nature and the position and velocity of every

particle in the universe at an instant, could predict and retrodict the detailed character of the world

at any moment of time. The 18th century astronomer‟s mind is an approximation of Laplace‟s daemon,

and much of today‟s „ mechanistic‟ analysis and design of farming systems reflects that thought.

Translated into today's reality it means, for example, that knowledge about DNA can secure „ food for

all‟ , or that diseases can be prevented by knowing a causative germ. Importantly, however, Newton

could only calculate the orbits of two bodies, and not of three or more interacting bodies where complex

„ relations‟ confused the regularity of the clockwork universe. Uncertainty was not overcome by Newton:

he „ shelved‟ it, and his deliberate simplification of nature was forgotten by his followers.

Indeed, Newton's followers took the „ two body‟ simplification as an article of faith, just as the dispute

about uncertainty between Einstein and Bohr, some two centuries later, was essentially an argument

about faith. Bohr accepted uncertainty as a physical phenomenon, while Einstein tried to furtherunderstand the „ logic‟ of God‟s thinking. He could not accept that God worked with chance, „ that God

played dice‟ , even though he accepted the uncertainty-relationship established by Heisenberg6 for its

instrumental value. These arguments reflect the difficulty of the church with Galileo‟s heliocentric

world-view and Kepler‟s ellipses. Cardinal Bellerminy could accept the formulas of Kepler and Galileo

as „ instruments‟ , as long as they were not meant to define a world-view, or to affect the existing

paradigm. Does it also perhaps reflect today‟s difficulty on the part of mechanistic thinkers in current

„ science‟ to accept that simple predator-prey models have more than merely instrumental value for

agricultural development. Do they have a „ religious‟ difficulty in accepting that „ man‟ is part of nature

rather than „ in charge‟ of nature? Where does modern thinking on agricultural development position

itself, and how does our choice of method and paradigm affect the outcome of our analysis and

prediction? Do we continue to look for certainty of measurement, or do we accept uncertainty and theneed for participation?

Since the 1960s the notion of uncertainty has acquired a much more prominent place in system thinking

(Klir, 1991). A further breakthrough in thought involving uncertainty of physical systems came with the

advent of computers and new insights from thermodynamic theory. And it was also recognised by the

founders of SSM in organisations, companies and communities, i.e., situations where humans play a

central role (Prigogine and Stengers, 1985; Gleick, 1987; Checkland, 1999). The existence of

uncertainty in both social and physical systems is hardly surprising if one remembers that the mind-

matter distinction is man-made, and that it has led to a situation where so-called „ sciences‟ and „ arts‟

seem to have lost the capacity to communicate. For some one hundred years now, the time span covered

by this book with respect to agricultural developments, the concepts of relativity and quantum physics

6 It states that one cannot know both the place and the velocity of an electron simultaneously.


5/27

5

have exerted their influence on the science of the very big and the very small. Now complexity has

started to influence modern system thinking in and around agriculture, but the debate on uncertainty and

on the validity of former distinctions is yet to show its full impact. Concerns about climate change,

recent outbreaks of BSE, foot-and-mouth disease, SARS, Avian influenza, political change and the

threat of war, computer viruses, scepticism with regard to globalisation, etc. demonstrate only too well

that „ one can expect unexpected things to happen‟ , and that values and paradigms can undergo suddenand dramatic change. We appear to have returned to the days of Heraclites, who stressed fluidity in

systems, in the days of Plato who talked of reflections of a real truth, and in the days of Aristotle, who

looked for patterns and drives behind systems. The challenge is to use these concepts beyond their

instrumental value alone, and to let them have an impact on our global view of agricultural development

in the 21st century.

Modern system thinking in agriculture

Modern system thinking concerning agriculture exists in different forms, which we here classify into

hard-, soft- and complex methodologies (HSM, SSM and CSM)7. Together they span the space between

systematic approaches on the one hand and systemic approaches on the other. A systematic approach

emphasises objective measurements, quantification, reductionist thinking and mechanistic synthesis. In

other words, and metaphorically speaking, the observer does not affect the „ clockwork ‟ , but „ he‟ knows

what happens anywhere if one part of the system is changed. In this approach, parts can be studied in

isolation and they can be engineered to „ control‟ the future. The work on DNA to understand the maize

plant‟s effect on a farmer's income is a case of mechanistic thought that resembles the simplification of

Newton's two-body problem. Basically, nothing is wrong with such thinking unless it pretends to

explain everything. A systemic approach assumes that the observer is part of the system through the

choice of parameters and methods made. It stresses change in and around a system, as well as the need

to include qualitative aspects of the mind in addition to „ hard facts‟ about matter. Indeed, systemic

approaches stress that, for example, the direction of agricultural development depends on whether the

researcher chooses to include non-physical aspects of farming and/or different perceptions about reality,

e.g. those of agro-industry, politicians or organic farmers. To state that choices about method affect theoutcome sounds like self-evident common sense, but it was precisely the notion that Einstein refused to

accept in Bohr‟s theories, other than for instrumental value.

One form of research on agricultural development is called farming systems research (FSR). It

originated in the late 1960s and early 1970s from work on the so-called Green Revolution, because

results from laboratory settings and experimental fields were manifestly not applicable in variable

contexts, unless the field conditions could be „ controlled‟ in terms of availability of water, new seeds,

chemicals and standard farming methods. Another reason for the emergence of FSR was that many

technologies were showing unexpected trade-offs due to operative relations both within and among

systems: a reflection of the three-body problem. The investment in FSR is impressive. It uses aspects of

HSM, SSM and CSM, and encompasses approaches applicable to both tropical and temperateconditions (Conway & Barbier, 1990; Collinson, 2000). For the purposes of this book, however, we

only elaborate on the characterisation of thought and paradigms in HSM, SSM and CSM while staying

away from the more practical and very useful procedures of FSR.

Hard system methodolo gy (HSM)

The initial HSM work in agriculture followed upon experiences from engineering and was characterised

by an approach expressed by De Wit in his inaugural address as:

7 Like other classifications, this distinction between HSM, SSM and CSM is man-made and of limited value,

but is deemed useful for the discussions contained in this book. Many other classifications are possible; theInternational Society of System Science has some 30 special interest groups, each representing a different

orientation in system thinking.


6/27

6

'…the development of models of agricultural sy stems (or of biological systems in general) has to

follow a heuristic approach that requires testing at two levels, i.e. the explanatory level, where

relations should be in accordance with observed phenomena at process level, and at the explainable

level, where model results should be in accordance with observed phenomena at system level' (De

Wit, 1968).

This mechanistic approach was especially useful in Western agriculture where farmers operated highlycontrolled production systems and where uncertainty was externalised through the use of external

inputs. The approach also appeared useful in identifying technical constraints in the agriculture of the

developing world, but the focus of much agricultural research was on the relative certainty of issues

concerning matter. As said by pioneers of the green revolution in Mexico: „we are concerned about and

aware of the complexities of social evolution, but […] we have restricted our discussions to subjects

[of matter] that we have studied intensively for many year s‟ . (Stakman et al ., 1967). Much valuable

work in this field was done in Europe, the US and Australia by scientists like C.T. de Wit, C.R.W.

Spedding, L.T. Evans and R.S. Loomis (see Rabbinge et al., 1990; Spedding, 1990; Evans, 1998 and

Loomis et al ., 1976).

Almost from the start, criticism was voiced concerning the 'technocratic' basis of this HSM and its

neglect of the 'human factor'. Indeed, in different parts of the world, and in many different disciplines,

the need for a „ different‟ approach became increasingly clear (cf. Checkland, 1981; Chambers et al.,

1989). In spite of these criticisms, however, it is to the credit of HSM that the world has seen two- to

three-fold increases in yields per unit area or per animal, in both tropical and temperate climates. The

sense of achievement and „ control‟ by HSM workers was expressed at the 75th anniversary of the

American Society of Agronomy by one of its former presidents:

„The foundation for much of the progress in agriculture […] has been laid by crop and soil scientists.

[…] The cultural practices, farming systems, fertilisers and other chemical inputs used in modern

agriculture are creations of your hands‟ (Brady, 1983).

The HSM thinking behind these yield increases was based on reductionism, emphasis on aspects of

matter and visions of a mechanistic clockwork universe. They define a system rather statically as being:a unit with well defined boundaries and a well-defined goal that consists of interdependent parts that

transforms inputs into outputs (Figure 4.1).

By implicitly focussing on „ useful‟ and measurable outputs, the HSM tends to ignore the „ mind‟ aspects

and the effects of „ waste‟ . Also, by „ freezing‟ a system into well-defined boundaries, it tends to

overlook variability and changes in time and space. It ignores a definition whereby a system is:

a way of doing things, an established procedure (Longman, 1985).

Hereafter, the „ static‟ definition refers to the structure of a system (we use the term „ form‟ throughout

this book). The static appr oach is also used by HSM to consider other typical “system” notions like

hierarchy, parts and boundaries (Box 4.1). Those notions are valid in all forms of system thinking, butthey are only applied in a rigid way in HSM. The „ dynamic‟ definitions refer to dynamic aspects of

system behaviour over time (called „ processes‟ in this book). We often refer to the forms and processes

of systems, and also to their combination that we refer to as the mode. Both form and process are in-

separable aspects of system behaviour, but a terminological distinction is retained in order to ensure that

neither aspect is forgotten, similar to the distinction between mind and matter mentioned previously.

Mode changes occur when form and/or process experience inordinate stress.

The static view of systems deliberately reduces interactions between the system and its context

(surroundings) while focusing on measurable „ matter ‟ entities, often embracing only the short term. It

echoes Galileo‟s emphasis on the need to ‟measure‟, from the time that mysticism and dogma‟s ruled. In

its extreme form, the system‟s description and boundary definition by HSM occurs in such a way as toconstrue the system as being affected by its context while the system itself does not affect the context. In


7/27

7

doing so, HSM marginalizes uncertainty, an approach that is known in classical economics as ceteris

paribus (everything else remains the same), and in Newtonian physics as the „two body‟ simplification.

This approach may be convenient and at times useful, but it erroneously assumes that a small effect is

no effect, an assumption challenged by the „ butterfly‟ effect described in the section on CSM. Thus,

HSM tends to conceive of a system as a temporary frame taken out of a sequence of events. It might,

for example, state that 1 kg of nitrogen from an external source yields 20-30 kg of grain, without payingattention to the long term side-effects of that nitrogen in terms of matter (groundwater enrichment)

and/or mind (farmer's income and social position). Typically, HSM does not specialise in the prediction

of trade-offs, such as emerging and inherent inequity between wealthy and less well-to-do farmers,

increased risk associated with continuous mono-cropping, environmental impacts of using large

quantities of chemical inputs, or decreasing efficiency of input use at high input levels.

Figure 4.1.A system as a unit, in this case an animal with input in the form of feed, fertiliser, etc., with

output in the form of „ produce‟ such as milk, meat, eggs, draught, and with „ waste‟ such as dung, urine

and heat. Particularly the long-term effects of waste were for a long time neglected in the reductionist

tradition of HSM.

Eventually HSM developed tools to address its own deficiencies, such as the analysis of system

behaviour at different spatial scales from farm household level upward (Box 4.1), and the effects of

different objectives and their trade-offs. Interactive Multiple Goal Linear Programming models are

typical examples of a HSM approach, which attempts to introduce the effects of bio-physical factors

and economic considerations in terms of technical relationships (Table 4.1). They explore trade-offs and

the scope for development at farm and regional level, but any conflict between different optimal

solutions, such as in Table 4.1, is left for society at large to resolve (Van de Ven, 1996; De Wit et al.,1988; Vereijken, 1997; Rabbinge and Van Diepen, 2000; Van Ittersum et al., 1998). Thus the

usefulness of these approaches is continually criticised. Uncertainty about the perceptions of various

stakeholders and policy changes undermine the validity of the HSM emphasis on „ objective‟

measurement and mechanistic relations. Both, SSM and CSM offer opportunities to understand system

behaviour beyond quantities, mechanistic relations, and aspects of matter alone.

Soft system methodology (SSM)The variety existing in the perceptions and „ goals‟ of stakeholders in different sectors and at different

levels of modern agriculture was basically sidelined in HSM, just as Newton‟s followers shelved the

three-body problem. However, this variation was encountered elsewhere at the start of the GreenRevolution, and also workers like Checkland (1981) found that existing assumptions of objectivity and

certainty in system analysis and in the design of commercial companies simply did not hold. They

coined the term „ soft system‟ , a somewhat unfortunate choice because it might be interpreted as

suggesting that soft interpretations are to be taken less seriously than those of HSM. SSM stressed the

point that employees or farmers have different perceptions of their own situation relative to those of

directors or policy makers. Moreover, the perceptions of both groups may change over time, for

example due to the effect of observers that come with questionnaires designed for policy makers, or due

to changing policies, prices and even jealous neighbours. Climate change, outbreaks of disease in

animals and humans, or events like those of the twin tower attack in September 2001 are only a few of

the many other „ incidents‟ that have changed public opinion and personal goals in the past few years.

produce

"waste"input


8/27

8

Table 4.1. The best extreme values of six goals optimised (bold) in the columns and the associated

values of other goal variables in the rows, all per hectare per year, for the experimental dairy farm „De

Marke‟ in the Netherlands (Van De Ven, 1996). Note that the optimum for one criterion does not tend to

coincide with the optimum for another criterion, a notion elaborated among others in Box 4.1.

Goal Milk prod.( 12000 kg)

Lab. inc.Dfl NO3

(kg N; 34) NH3

(kg N; 30)P surplus

(kg P; 0.5) N surpluskg N

Milk prod. 17660 3700 34 30 0.5 225

Labour

income

15350 4380 34 30 0.5 269

NO3 leaching 12000 1650 13 22 0.5 165

NH3 volatilisation 12000 2145 34 17 0.5 160

Total P

surplus

15180 4210 34 30 0.0 269

Total N

surplus

12000 1120 24 20 0.5 94

Box 4.1. Hierarchy, common versus individual interest, and fractals

Individual systems do not exist in isolation from each other. They interact and together form larger

systems in a hierarchy where boundaries are hard to define or even non-existent. For example, several

cells (made up of organelles) make one organ (root system, kidney, brains). The organs together make a

larger system (plant, animal, man), they together make farms, village communities, regions, etc. Each of

these levels can be considered as a „ separate‟ system, i.e. it is always important to state at which level

one is working (organelle level, organ level, and so on to national level and beyond). Cells themselves

are infinitely complex, but their complexity simplifies itself as one moves one level higher in thehierarchy where one considers a cell without „ losing the wood for the trees‟ in the details of interaction

at individual cell level. This also implies that for a larger system to function, one needs to accept that

the lower level systems adjust themselves towards the larger whole, but that the reverse is also true!

Farmers tend to be interested in plot, herd and whole farm yield, not so much in the specific individual

yield of a plant, animal or other component. Governments tend to be interested in the „common‟ well-

being or gross national product of a region, if necessary at the expense of certain individuals.

Not only does the complexity of systems at lower levels simplify itself at higher (and lower!) levels.

Systems also tend to repeat behaviour at different levels of space and time. This aspect of repetition is

reflected in the notion of fractals. For example, measuring the length of a coastline appears to be easy if

one has a given „ yardstick ‟ that overlooks irregularities like rocks. However, with a finer „ yardstick ‟ one

encounters new complexities, such as how to factor cracks and irregularities in rocks into the measuring

process. This metaphor was developed by the mathematician Richardson in the early 20th century. It

was revived by Mandelbrot in his „discovery‟ of fractals, mathematical reflections of real life

phenomena where systems at several levels of space and time repeat similar behaviours. For example,

irregular feed intake in animals over the day is matched by irregular feed intake over weeks, seasons,

years, countries; irregular energy fluxes and nutrient supply in plants, or farms and regions show

similar patterns. In the same way, one can better understand the principles of mixed farming by learning

from similar processes at cell level, plot level or international level. A paradigm shift is implied by the

realisation that system behaviours tend to repeat themselves whether „ we‟ are cell, „ man‟ or rabbit and

fox populations. All of a sudden we are part of nature, not above it, in spite of the special

responsibilities and choices that we may have (Cohen and Stewart, 1994; Schiere et al., 1999;

Mandelbrot, 2000).


9/27

9

It can indeed be said that the uncertainty factor likewise exists in agriculture, people, animals, and

crops, and also that, crucially, farming systems are learning systems. Indeed, many consumers and

farmers have started to question the control attitude of modern approaches to agriculture, which in their

view results in the exploitation and/or destruction of natural resources, or in unequal access to

resources. Sceptics have begun to question how long we can continue to use fossil reserves at currentrates, or why most of the land should be in the hands of a proportionate minority. Some farmers turn to

organic farming while others hesitate to make investments necessary to remain apace with recent trends

in farming. In other words, „ attitude‟ and perceptions are questioned, while previously accepted linear

„ hard‟ facts and traditional paradigms are reassessed. Consequently HSM struggles to follow, whereas

SSM takes changes of mind and „ ethical‟ considerations as an integral part of life. SSM internalises the

concept that the observer's choice of measurement (paradigm) affects the outcome of the research. The

Einstein-Bohr argument about uncertainty in physics becomes apparent here in decision-making about

farming as demonstrated by the differences between HSM and SSM. In addition, SSM stresses that a

cow or a field, for example, is not just a physical entity. The cow or the field also has emotional value

(meanings, goals) for farmers and consumers, a notion that is captured in the term „second order

information‟ (Ison & Russell, 1999). Thus the distinction between matter and mind becomes less clear

or even counterproductive; it is a man-made distinction after all. The concept of „ learning systems‟ from

SSM has more than merely an instrumental value. It is a departure from the HSM notion that systems

have an explicitly defined goal, rather than continuously changing goals. It also requires a paradigm

shift in terms of the method used for choosing policy in agriculture and in daily life, towards a new

balance between control (HSM) and participation (SSM), as expressed in the Spanish and Indian quotes

at the beginning of this chapter.

The inherent participation (interaction) of the observer in the act of observation is illustrated in the

fishnet-metaphor by Addington (quoted by Ashby, 1958):

„An empirical scientist who threw a net into the sea examined the catch, and then announced the

empirical law that “all sea creatures are more than two inches long.” ‟

This metaphor re-introduces ethics and choice into system analysis by stressing the now familiar pointthat method chosen determines the outcome of research, since no combination of measurements can ever

give the full picture of reality. The fishnet metaphor is paralleled by the coastline metaphor mentioned in

Box 4.1. Essentially it states that the length of the coastline increases as the scale of measurement

decreases. Eventually such measurement takes into account also the shape of sand grains and cracks in

rocks. Ultimately one ends up „ measuring‟ the circumference of the rocks and grains, or even molecules

or intercellular cavities. This effort of obtaining exact measurements makes one lose sight of the original

question, a well-known experience in much objectivist and reductionist science. To measure is „ to know

only very partially‟ , full control is an illusion, choice is inevitable, and the learning process forces the

observer to continuously adjust the monitoring approach. In addition, the measurement intervention

itself determines the outcome, a possibility that is marginalised in HSM, but stressed in SSM whose

followers are well aware that their method of measuring affects the answer. Ultimately SSM challengesthe partly Aristotelian notion of a „ goal‟ . As stated by Checkland and Scholes (1990):

A system does not have a goal but it is given one according to its context (and by implication:

changing contexts result in changing goals; added by JBS).

Furthermore, SSM is not focused on precise definition of system boundaries; Röling (1994) says:

A system is a construct with arbitrary boundaries for discourse about complex phenomena to

emphasise wholeness, interrelationships and emergent properties.

The idea of changing goals and vague system boundaries bewilders some HSM practitioners. It does

justice, however, to uncertainty and the existence of unstructured situations in „ real life‟ that is full of

change and contradictory perceptions. For example, a boundary for a cow is not a boundary for a

nitrogen molecule. And a boundary for „ matter ‟ may not be a boundary for „ mind‟ , just like the timescale for a farmer tends to differ from that of a distant policy-maker who studies macro-economics. Use

of SSM implies that the observer accepts the existence of different world-views. It is a paradigm shift


10/27

10

that goes beyond instrumental value: SSM explicitly restores the uncertainty as well as the active role of

the observer (stakeholder) within the process of enquiry and in coping with problems. Participation and

adaptive management, defined as continuous (learning) response to feedback becomes necessary in a

worldview where systems change with the context (Chapter 10). The Landcare program in Australia

and environmental co-operatives provide examples of what participatory approaches can do if policy

permits: rural communities address local problems according to local priorities (Chapter 6).

Complex system methodologyIssues of system dynamics and continuous learning are sidelined and shelved in HSM, whereas they are

part and parcel of SSM, and they occupy a central position in CSM. For HSM workers the hardest part

of the paradigm shift required to adopt CSM is that they have to start dealing with the uncertainty

associated with changing and multiple perceptions, a notion already present in SSM. One might say,

however that SSM workers on their part find it hard to adopt CSM because they tend to distrust the

mathematical and physical origins of CSM. The SSM tradition is perhaps just as stuck in the „ mind‟

mode as the HSM tradition is stuck in the „matter‟ mode. One „ novelty‟ of CSM is that it extends

uncertainty accruing from human systems into physical systems and vice versa, thus bridging the man-

made gap between matter and mind (Capra, 1997). It goes beyond the dynamic models used, for

example, by Meadows et al ., (1972). It stresses the emergence of real mode changes and completely

new systems, not extensions of previous ones (Chapter 5).

Any precise definition of complexity would be an oxymoron, but a characterisation of a [complex]

system from a CSM point of view could be:

A complex system has innumerable emergent properties, hard or even impossible to define

boundaries, and relations and characteristics that are open to an infinite number of different

interpretations.

Characterisation rather than precise definition is inherent in the notion of uncertainty. It has the

advantage of providing flexibility but, conversely the disadvantage that one loses control. Another

implication of this characterisation is that results cannot be traced back to single causes. In terms of thediscussion on biotechnology, DNA modification only results in higher yields and/or farm income (these

two are not necessarily related) if „ allowed‟ by a context consisting of climate, farmers, soil type and

pest/disease pressure. Basically, this is Newton‟s simplification of the three body problem revisited.

Control methods to trace the „ real‟ cause can be powerful, but they eventually fail due to administrative

detail and high costs, subjective choice of boundaries, or by overlooking relations between a system and

its surroundings. Systems do indeed affect their context in a way that eventually leads to unexpected

dynamics. Typical examples of such unexpected effects are the social costs associated with the outbreak

of foot-and-mouth disease in Western Europe in 2001, the political consequences of the SARS outbreak,

or the social change associated with the plant breeding programme underlying the Green Revolution.

Both higher yields and disease affect the mind as much as the matter of the farming community.

A landmark contribution to the theory behind CSM was the chance discovery of the „ butterfly effect‟ ,

by Lorenz in the 1960s (Gleick 1987). He found that predicted complex (weather) patterns changed

dramatically, with (even minute) changes in the initial values of the model. Butterfly effects may get

dampened and cancelled out by other effects (boundaries!) They may also combine with other processes

to become stronger, a typical case of interaction between systems and their context. One example was

the introduction of rabbits in Australia (a whim in somebody‟s mind) that combined with an absence of

natural predators to cause environmental disaster in biophysical and socio-economic terms. Other

examples are the chance emergence of a prion that causes BSE in a world with cheap transport and the

recycling of animal products: the beef scare is a „result‟ and a cause with yet unforeseen consequences!

Lorenz‟s butterfly effect led to the uncertainty factor entering the domain of matter, i.e., in ma thematics

and physics. Uncertainty could no longer be confined exclusively in the mind domain, because nothingcan be measured more accurately than the small deviation that is large enough to cause unexpected

behaviour. Basically, it has to do with the notion contained in Gödel‟s statement that:


11/27

11

„if you fix the rules of inference, and any finite number of axioms, there are meaningful statements

that can neither be proved nor disproved‟ (Ruelle, 1991).

The CSM tradition starts by recognising that important factors underlying unexpected system behaviour

include notions such as::

The butterfly effect: uncertainty about the initial conditions (see the assumptions by Laplace in the

quote at the start of this chapter). The „lock -in‟, i.e. positive feedback and the delay between action and reaction. This occurs, for

example, when a society is stuck with its infrastructure, or when a business is stuck with an

investment or attitude that prevents adaptation to changing contexts (Box 4.7).

Non-linear system behaviour as a consequence of the fact that a given system‟s rate of growth

depends on the context and condition of that system at the preceding moment. For example, the

growth of an industry, or plant or animal population, depends on its size at a preceding moment and

on the past, present and expected resource flows (Box 4.6). The„ learning‟ of a system depends on

its previous experiences; relations exist between past and present contexts and systems.

Relations between sub-systems as well as between systems and their contexts (Box 4.2). A change

in one part affects other parts, that is the holistic aspect of complex system thinking going beyond

Newton‟s „ two-body‟ simplification. For example, the decision of one farmer to plant wheat maydepend on others deciding to do the same. And new varieties or cropping patterns eventually affect

ground water levels, social organisation, and so on, well into uncertainty.

The relevance of such notions for policy-making is that rigidity and emphasis on administrative

measures for system control are unlikely to be effective in the long term. Beyond certain (threshold)

values, systems start to respond by behaving differently. They yield less, or they diversify into different

modes as implied in Box 4.2 and Box 4.3. Complex problems that have no easy solution tend to result

in different coping strategies, the combination of which results in the next phase. More refined

measurements and extrapolations may be counter-productive if they obscure other avenues. They may

even „waste‟ resources that could better be used for more relevant and/or emerging phenomena

elsewhere if they keep us locked in tradition („ lock-in‟ ). For example, the post-World War II emphasis

on high yields in Australia and the Netherlands ignored the emerging problems of salinisation and

eutrophication. And the recent outbreak of foot-and-mouth disease in Europe illustrated two other

potential difficulties with the linear measurement and control approach. First, a centralised rigid and

large bureaucracy can aggravate an emerging problem. Second, one sided emphasis, ostensible control

of „ accidents‟ , may actually lead to higher social and physical costs, rather than to more efficient overall

production.

The „ emergence‟ of a combination of several coping strategies, rather than one or a few isolated

solutions is central to CSM and it is a major characteristic of non-linear system behaviour (Box 4.3) .

One of the coping strategies is the decay or „ death‟ of an existing system, a „ solution‟ that is politically

often hard to accept. However, it tends to allow for the emergence of new forms. Schumpeter calls this

„ creative destruction‟ (Holling, 1995). It is a notion remarkably close to Hindu thinking on creation anddestruction operating in parallel. Another aspect of working with a combination of coping strategies is

that one „ allows‟ a system to choose a particular mode out of several alternatives, depending on what

other systems do. This introduces a variation of forms and surprises (diversity and serendipity) that

eventually leads to unforeseen consequences and continuous dynamics. The notion of „solution‟ rather

than „coping strategy‟ tends to imply only one choice with a final and static result. Too often we tend to

train our students with just such a notion of solutions by offering them „ simple‟ problems (Schumacher

1972). This leads to one-problem-one-solution approaches, and to linear efforts being undertaken to

maintain growth in, for example, a given type of food production. Such an approach is likely to

eventually strain systems and their contexts to such an extent that they encounter limiting factors and

crises (see Chapter 5 and below in this chapter). In real life it is likely that unexpected combinations

occur and that systems diversify (Box 4.3). Dutch farmers have responded in various ways to a series ofemerging problems, such as overproduction, environmental impacts, animal welfare, etc. These coping


12/27

12

strategies include emigration, buying additional manure quota and/or more land, hoping for better

prices, sale of home-products, small-scale tourism on their farm, quitting, etc. Much linear policy-

making and teaching tends to shelve small creative efforts as not useful for „ the sector as a whole‟ , or it

continues to seek prototype „solutions‟ that can be generalised. „ Chaos‟ -management is keen to identify

new exceptions and sees diversity as an opportunity. Creativity is an essential ingredient where

mechanistic approaches fail, and „mess‟ is important when mode changes cannot be avoided atreasonable cost.

Box 4.2 The importance of context and the inseparable nature of system and context relations.

Context and the active interaction between context and system is pervasive to such an extent that the

distinction between context and system becomes blurred. Context acts, for example, where a bee egg

that is well fed („ good‟ context) develops into a very reproductive queen, while the same egg in the

context of „ bad‟ feeding develops into a sterile worker. In other words, the information in the egg

expresses itself differently depending on context. It is not „nature or nurture‟ but „nature and nurture‟,

nor is it „external or internal‟ development. It is the interaction between the system and its context (if

those concepts can even be used in the traditional sense). The same holds true for the seed of a pine tree

that develops into a crooked tree once it germinates in an open space (provided it is in a context with

enough water, no fire, etc.). In a forest, however, the same seed develops into a straight tree, not after

first trying to grow crookedly, but directly at the first attempt. Quite different and similar in this

connection is the mode-seeking in weather systems: depending on a particular combination of relative

humidity, temperature and air pressure, clouds will form, all of the same general shape that belong to

that particular type of climate. Other cloud shapes would be formed under different temperature and

humidity conditions. Similarly, a farmer operating in densely populated areas has a different view on

farming than one operating in the far outback; matter and mind are linked, system behaviour and

context are linked.

A fascinating example is the case of the fertilised egg in a conducive context, as shown in the above

diagram (I). It splits into two similar ones (II) and so on until the cluster of eggs consists of cells with

similar genetic information but different self-induced contexts (IV-VI). Untill stage III all cells have a

similar context, though different from their predecessors. After stage IV the „inside‟ cells, however,

have different context than the „outside„ ones. Somewhere at this stage, the „whole‟ system starts to

diversify into what could be called epithelium and endothelium cells. Such development has a strong

parallel in the development (or co-evolution) of farming systems (Box 5.2). System development in

this CSM notion is the result of dynamic relations between system and „its‟ partly self -generatedcontext. The HSM notion of a „ frozen‟ system with a fixed context of ceteris paribus needs to be

reconsidered if one takes these concepts for more than instrumental value alone. A particular form and

process of farming may emerge due to a given context, but sooner or later it is also the farm and farmer

that determines the context, and then back again, and forth again, but always onward.

The butterfly effect was one of the discoveries that undid the clockwork notion and re-introduced

uncertainty. It matches other work from the second half of the 20th century concerning complexity, also

known as non-linear system dynamics and „ chaos‟ (Klir, 1991). We use the terms complexity and non-

linear thinking to stress that this refers to a rather young and not yet well formalised approach. The fact

that a change in one part of a system may have an effect elsewhere starts to be hesitatingly accepted in

HSM. It is present in the concept of learning systems in SSM, and inherent in the notion of trade-offs

(Conway and Barbier, 1990). It is central to the complexity thinking of CSM which states that complex


13/27

13

problems have more than one coping strategy (Box 4.3), that several perceptions exist (Box 4.4), that

uncertainty is the rule, that certain modes are more likely than others and last but not least, that system

and context are inseparable and yet different. Interrelations lead to trade-offs, their existence has more

than merely an instrumental significance. Trade-offs and loss of control therefore become part of our

world-view. Efficiency gains in one place are likely to cause efficiency losses elsewhere. The optimal

use of limiting factors, the timely shift into other modes and, as a matter of ethics, the judicious andrespectful use of finite resources and nature then become relevant. Trade-offs occur at and across all

levels of system hierarchy (Table 4.1) as illustrated by the following examples:

At crop level, emphasis on total plot yield tends to occur at the expense of individual plant yield.

At animal level, breeding for breast-meat in turkeys resulted in physical limitations in terms of

reproductive behaviour.

At flock level, attention to high laying percentages in hens implies the „ emergence‟ of male chicks

that have to be destroyed due to inferior meat production characteristics.

At farm level, profitable forms of mono-cropping may lead to more disease pressure or soil erosion.

At regional level, a dam for irrigation downstream may displace local populations upstream.

Box 4.3 A one-problem-one-solution approach versus use of combinations of coping strategies

Teaching is often based on questions and problems that have one solution (1+1=2). Some problems,

however, have more than one solution (4 = 2) and complex problems tend to have no solutions.

Taking an example from animal nutrition: it is impossible to mix two feeds, A and B, with 50% and

70% TDN (an energy value) respectively and 10% and 20% crude protein in such a way as to arrive

at a mixture with 60% TDN and 18% crude protein. Such „ impossibilities‟ are common in examining

the trade-offs between higher agricultural production and lower resource use, or more „ nature‟ value

(Table 4.2). A strange paradox in non-linear thought is that an infinite number of coping strategies are

available once we accept that a particular solution is impossible (or even if we accept that the solution

has an undesirable cost). The choice of the observer once again becomes an operative factor e.g. in the

decision whether or not the benefit of a „ solution‟ outweighs its cost. For example, in the case of thefeed mixing, it is possible to achieve an infinite number of feed mixes from A an B if they both contain

> 60% TDN and > 18% CP.

A coping strategy is a way of dealing with a problem in the knowledge that a „ perfect solution‟ does

not exist. One gains something while losing something; your gain may be my loss, or vice versa, some

prefer this, others prefer that. In accepting the existence of trade-offs, one should ask which set of

options is available for different farmers to cope with a crisis, including the option to quit. Complexity

and chaos management stresses that more than one option is available, none of them is 100% perfect

for all stakeholders, and each one leads to continued system dynamics (co-evolution). The change from

thinking in terms of one solution towards a combination of coping strategies requires a paradigm shift

from static towards dynamic thought.

Not all feed-backs or trade-offs are negative, indeed at first they tend to be „ positive‟ , but carried too far

they tend to „ produce‟ „ negative‟ side effects, an essential feature of non-linear system behaviour.

Whatever the situation, the choice of criteria and trade-offs determines the analysis and design of new

systems; the observer affects system behaviour, man and farming are part of nature. Tensions in

farming do not just happen, they are part and parcel of development and they occur at all levels of

system hierarchy. In the CSM world-view, the image of agriculture being marred by „ problems‟ is not

an accident, it is the consequence of a combination of different perceptions, past choices, non-linearity

and the unpredictable effects of „ God playing with dice‟ . Continuous alertness to feedback is required,

while the need for participation overtakes the notion of mechanistic control. Managers, policy-makers,

teachers, researchers and farmers have choices to make and signals to interpret; learning systems are a

necessity, not a luxury. So is there any certainty at all?


14/27

14

Box 4.4 Perceptions of fluctuation and stability/resilience by different observers

Different perceptions on the severity of fluctuations in a system reflect an observer‟s choice for the

scale of observation. For example, a farmer having a time horizon for survival between points A and

B will experience irregularity, but an administrator dealing with a time horizon scale from A to C willnote regularity! This issue has been worked out, for example, in great detail for perceptions about

sustainability and size of catch in fish populations by Van Densen (2001).

Form and thermodynamic theory, a static approach.

CSM and „chaos-theory‟ stress a strange tension between repetitions of form on the one hand and on

uncertainty on the other. Repetition of form can be referred to, among others, as „ fractal behaviour ‟ , a

term that is related to the concept of fractals. Fractals represent models of systems that repeat

themselves at different levels of system hierarchy (Mandelbrot, 2000; Gleick, 1987; Capra, 1997). For

example, horses, cows and frogs may all be different, but they all have eyes, kidneys, and digestive

systems. Also, trees and bushes are different, but they also have many similarities in terms of form and

processes, depending on how time and space scales are defined. Important, not all similar forms need to

stem from a single original such as expressed in Plato‟s „ idea‟ and/or Sheldrakes „ morphogenetic fields‟ .

The eye has emerged independently several times in the course of evolution; similar clouds form without

mutual „ knowledge‟ of each other; different farmers find similar solutions without consulting each other.

Processes such as exchange of resources and mutual adjustment of parts occur in the cell, the organ, a

plant, a plot, a farm, a village, and at regional and international levels. Animals and plants all require

nutrients and energy, while they excrete waste and dissipate resources, and so do humans, societies and

cities. Genotype-environment interactions are not unique to animals or plants, they are expressions of

general non-linear system behaviour. Time and space scales may differ, but principles are often similar

(Box 4.1).

CSM stresses another apparently paradoxical strange tension, i.e. between a focus on uncertainty and aconcomitant stress on the general validity of two laws of Thermodynamic Theory (TDT). The latter can

be assumed to determine form and behaviour of all systems, independent of their size. The laws have not

been conclusively proven – another strange uncertainty – but their validity can be described as being

beyond reasonable doubt. They state that:

First: energy and matter cannot be created nor destroyed, it can only change form.

Second: a closed system left to itself tends to greater disorder, expressed in terms of increased

entropy.

Often, people from the „ mind‟ orientation in SSM find it hard to accept that laws of „ matter ‟ affect our

perception about life but they should remember, like their „ matter ‟ colleagues of the HSM, that the

mind/matter distinction is a man-made construct. SSM thinkers in this respect may take comfort fromthe notion that TDT reflects system behaviour at a level that cannot be claimed exclusively by „matter‟

A B C


15/27

15

thinkers. Also, cornucopian thinkers of both the HSM and SSM tradition in agricultural development

find it hard to accept this notion of entropy and finite resources. They have a world-view in which

human ingenuity can overcome the limitations of nature, and where technology is a gravy train (Chapter

5). This section therefore discusses some aspects of emergence and repetition of forms, together with

other implications of TDT for farming.

Emergence of formThe first law of TDT implies that grass can be made into milk; that water, solar energy and nitrogen can

be turned into grain, and that energy stored in coal or wood can drive steam engines or keep homes

comfortable. It also allows the reverse to take place. However, the second law states that form tends to

disappear in closed systems which „are left to themselves‟ . It implies that it is harder to make grass

from butter than vice versa. It therefore also implies a notion of irreversibility, as observed for example,

when both hard and soft system thinkers find it hard to break certain acquired routines. Too often

farming systems, farmers, researchers, and policy-makers of all disciplines become locked in traditional

modes. They thus refuse to adapt, i.e. refuse to learn. It is a form of lock-in and delayed feedback that

is, for us, a major precondition for unexpected system behaviour.

Another important point is that the second law of TDT, when applied to nature, refers to open rather

than closed systems (Bertalanffy, 1968). All systems in agriculture and society are open, because they

receive a continuous but variable flow of resources that represents different forms of energy. They are

thus not „left to themselves‟ and when the resour ce flux is high enough there is a tendency for local

order to appear (Prigogine & Stengers, 1985). Systems self-organise according to the resource flux but,

in terms of the second law, this order has a trade-off in the form of greater total disorder elsewhere,

which may be hidden from view. For example, the use of petrol to plough a field seems a small price to

pay for the re-ordering of the field, but the disorder at molecular level, created by combustion, is much

greater than the order created at the level of our perception of reality, the field (Odum, 1971). Indeed,

according to the TDT a choice for order in one place implies disorder elsewhere. The „ mode‟ of order

depends on an interaction between the system and its environmental conditions, i.e. on the resource flux

(Box 4.5). Seen in these terms, a greenhouse industry in the middle of Australia is less likely to emergethan an extensive sheep ranch. And a potato crop is more likely to emerge in a Dutch polder than in the

Australian outback, unless „control‟ through enforced resource fluxes is applied from outside. It is a

challenge for policy-makers, farmers and consumers to appreciate the implications of the second law of

TDT and to match external „control‟ with the requirements for sustainable farming. It is another

challenge to match the demand of growing populations with realistic expectations regarding access to

resources.

Box 4.5 Form and the water-boiling metaphor where the mode depends on the environment, i.e., the

resource flux.

The molecules in a vessel of water on a gas burner will first continue to tumble about at random. The

flame adds energy as a form of external control to the vessel. The flame also results in disorder ofcombustion gases and movement at a molecular level outside the vessel. Inside the vessel the movement

of molecules is at random; an „equilibrium situation exists while the supply of energy from the burner is

negligible. Small additions of energy can be dissipated without appreciably affecting the organisation

(mode) of the molecules. However, beyond a certain point the water boils, molecules become clustered

into bubbles. In short, beyond a certain input the equilibrium is disturbed, systems self-organise and

new structures appear, some being more likely to occur than others. Square or triangular bubbles are

„ less likely‟ to occur than round ones, due to laws of nature, and given „reasonable degrees of outside

control. In the same way, it is unlikely that sustainable farming systems will be encountered where

animal or crop density exceeds the carrying capacity of the resource base (given reasonable levels of

system control).


16/27


17/27

17

changes) to other niches, rather than to overall efficiency increases. For example, cornucopian

coping strategies gamble on abundant availability and increased use of inputs to meet a specified

target. It is a typical HEIA approach that accepts mode changes and dependence on other resource

flows as a sign of progress. Other farmers, however, tend to cope by reducing inputs. They

participate with „ nature‟ by accepting that the quality of their natural resource8 base governs their

form and process of farming.

Box 4.6 The algae principle and niches for organisms and farming systems

The algae principle shows that a minimum input of energy is required to maintain the rather simple

system of the prokaryotic blue algae. Increased resource fluxes lead to a curvilinear response that

eventually results in less than maximum output. However, a more complex organism like the eukaryote

green algae in this case, needed a higher resource flux for maintenance, but it also produces more

(Elenbaas, 1994; Schiere, 1995; Allen, et al. 1999). The diagram shows that a certain input (X-axis) is

required to maintain the form of a given system, to offset or to compensate for the tendency of local loss

of form or increased entropy, as implied in the second law of thermodynamics. Initially, the output (Y

axis) increases as the inputs increase. This continues rather linearly up to the point where the organism

cannot cope with more input and where non-linearity starts to show. On proceeding from left to right on

the X-axis the resource flows increase. In that case another system with higher maintenance

requirements can better use the resource flow in that „ niche‟ to yield higher output, and so on. The

algae-fractal repeats itself in animals (here cows), and with „ improved‟ plant varieties, for example, as

resource flow increases due to increased soil fertility.

Real life systems often display more irregular behaviour than the stylised response illustrated above.

For example, organisms with a higher maximum efficiency than those illustrated in the diagram, would

lead to a different tangent. However, higher efficiency is always restricted to a particular niche (in other

niches they would have lower maximum efficiencies) and systems that depend on higher quality fluxes

to give high outputs may not be the most efficient.

8 The very notion of a resource base is even laden with a “choice” .. is it the resource base of humans, or are we

part of nature that can also consider “us” as its own resource base

output

α Input

Zebu cows Crossbreds Turbocows

on straw on grass on fodder / grains

Millets / rye Sorghum / barley Wheat / rice

on oor rain fed soils on better soils and hi h in ut conditions


18/27


19/27

19

The predator-prey relationPredator- prey cycles are „models‟ of system behaviour that occur in several forms. They can be depicted

in different ways, necessarily simplifications, but they have wide applicability, also for agriculture

(Holling, 1973). The essence of the predator-prey relationship is that the growth rates of predator and

prey populations are dependent on their previous size, they are interdependent, non-linear, and not in phase because of a lag in the growth-response of one population to the other (Box 4.8). In other words,

they fulfil the conditions for non-linear chaotic system behaviour as previously described. They also

challenge the paradigm of the mainstream world-view in at least three different ways:

Linear and cornucopian thinking tends to assume that continuous growth and/or control can be

achieved through the technological gravy train and/or an expanding resource base (Chapter 5).

Conversely, the conservationist world-view argues that „ predators‟ , such as humans, have to adjust

their consumptive behaviour to avoid collapse of the resource base. Both paradigms accept the need

for change, they differ in their thinking about access to, and need for, conservation of resources.

Classical economics tends to think that the ratio between rabbits and foxes will eventually stabilise

in a dynamic equilibrium. In other words, it assumes that wars against hunger can be won and that

„ solutions‟ are possible. Non-equilibrium thinking assumes that predator-prey ratios will not

stabilise and that every coping strategy leads to its specific trade-offs and inherent dynamics.

Both SSM and CSM stress that there are several coping strategies when predator populations start

to encounter stress situations such as in point t4 in Box 4.8. Standard or „ average‟ solutions are

exceptions rather than the rule. For example, the predator can „ select‟ from a combination of the

following strategies:

o reduce numbers through migration (Australia accepted immigrants from the land- starved

Netherlands) and/or by fighting each other

o continue to eat and feast while hoping for the best - that the „ dice‟ will come their way (the

Dutch discovered large reserves of natural gas in the 1970s, thus releasing resources for new

infrastructures)

o increase access to the resource base, such as through globalisation and use of resources from

elsewhere (see Box 4.9), or

o modify life styles, e.g. by going organic or by gambling on technological change. In other

words, change from one fox-type into another.

Such a series of coping strategies involves ethical choices and challenges for policy setting. In terms of

predictability, various options may exist, but the success of one is often determined by what the others

might do - „if he leaves, I can stay‟ . Control measures aimed at permanent solutions are counteracted by

system uncertainty and the inherent dynamics of learning systems in the broadest sense. Diversity in

both ideas and modes of farming is important because it provides options to cope with the future.


20/27

20

Box 4.8. A simple representation of a one-predator/one-prey relationship

The phase-space diagram below, also called attractor by „ chaos‟ workers (Gleick, 1987), is a simple

illustration of a one-predator/one-prey relationship in dynamic equilibrium. It shows how a rabbit

population (the prey) can increase rapidly from t0 to t1 because at t0 there are only a small number ofrabbits, a relative abundance of food, and few predators. The predator population (foxes) only starts to

grow after the prey population has started to grow. The number of rabbits continue to grow from t 1 to

t2, but this growth rate decreases because of increased pressure from predators and/or reduced food

supplies. When the rabbit population remains stable from t 2 to t3, the fox population continues to grow,

even from t3 to t4 when the rabbit population starts to decline. Subsequently, the rabbit population

declines further, leading to weaker foxes, and after some time the rabbit population finds itself growing

again with renewed abundance of food and low predator pressure.

Real life situations that resemble this type of model are the decline in quality of resource bases due to

exploitative ways of farming and/or consumer behaviour. Degraded sand dunes are part and parcel of

Dutch agricultural history but occur around the world more generally. Even the disappearance of „ clean‟

soil and air as a result of excessive „ waste‟ from fertilisers, animal manure, car exhausts and (agro-)

chemicals is likely to force the predator population that is called society into other modes. Like all

models this one simplifies but illustrates real system behaviour.

Holling's adaptive cycleHolling expanded on the predator-prey model by describing mode changes as resource supply fluctuates,

partly due to the system development itself (Box 4.9). For example, farmers moving into virgin areas

(represented by the top left quarter) can make a „quick buck‟ by rapid expansion though not necess arily

efficient exploitation of resources (bottom left quarter). However, after some time resources become

relatively scarce due to increasing numbers of farmers. In that case the efficiency becomes increasingly

important, and mode-changes start to make sense if they conserve resources by recycling within and

between systems. This process is apparent in Dutch farming of the past few decades. After WWII, the

Dutch agricultural sector had increasing access to resources due to increased access to world markets.

High, but „inefficient‟ modes of production developed: specialisation and highly productive systemswere the order of the day in the 1960s and 1970s. After some time, however, farmers started to face a

crunch that forced them into self-inflicted mode change. Resources were still abundantly available but

they could not be used as before, due to the difficulty in disposing of produce and waste products.

Dutch farming moved to the conservation mode of the top right quarter. It combined individual coping

strategies, such as improved efficiency, organic farming, precision farming, off-farm income, added

value and recycling between farms. As an important aside, mixing occurred at farm level till the sixties,

disappeared in the seventies, and began recurring at regional level „ between farms‟ in the late nineties, a

case of fractal behaviour that shifts over system levels.

Eventually, however, efficient conservation modes can become rigid, a nasty but fundamental trade-off.

Slight disturbances may threaten the whole network of interrelated subsystems. Collapse in such asituation occurs due to external changes and/or disturbances from within. Bankruptcy of a supply

company or the emergence of a virus can be the straw that breaks the camel‟s back. Pent up energy is

t0 t1

t2

t3

t4

rabbits

foxes


21/27

21

released and reorganisation takes place. The cycle starts again with new but often similar forms and

processes while Schumpeter‟s „creative destruction‟ recurs. A major world-view question here is

whether policy setting, at whatever scale, should aim for the single „ solution‟ of becoming a more

efficient (and rigid) system part-way through the conservation quarter (Box 4.9). Or should we also

„ gamble‟ for an array of coping strategies. Again, the importance of diversity cannot be stressed enough

for efficiency and survival. Unfortunately, and too often, the „locked-in‟ ancient regimes tend to hamperrather than to stimulate innovation, mostly from lack of will and methods to cope with change.

Box 4.9. Holling‟s adaptive cycle (Holling, 1995; Gundersen and Holling, 2002 )

The adaptive cycle is most easily understood by entering the „ pretzel‟ at the top left, a mode that tends

to occur after system collapse and/or sudden influx of resources ( e.g. after a flood, volcanic eruption,

disease outbreak or political instability). At that point there is a relative abundance of unused resources

that is captured by fast but not necessarily efficient colonisers (the „cowboys‟) in the bottom left

quarter. As colonisers multiply resources become increasingly scarce, hence collaboration and resource

exchange become advantageous. As a result, the more efficient but rigid systems tend to take over into

the top right quarter. Over there, access to resources declines for individual cowboy type organisms,

and connectedness increases, until the system breaks down again (the bottom right quarter).

Importantly, the release and reorganisation phase may take much less time than the exploitation and

conservation phase, as indicated by the different densities of the arrows. Moreover, systems will run in

different modes at the same time and at different levels of system hierarchy: a plant or animal itself is

quite interconnected and can store as many resources as possible, while organisation at the plot or herd

scale may still be relatively „ inefficient‟ . Mode changes occur at the end of each quadrant and

throughout, as well as at different levels of system hierarchy.


22/27

22

Concluding comments

Agricultural systems are subject to mode changes in time-space. This is due to the interaction between

their own development and (self-inflicted) changes of the environment, e.g. where population pressure,

resource-exhaustion or overproduction ruins the system‟s resource base. Other mode changes are„caused‟ by purely external factors, for example new export regulations, the emergence of new

products like (bio-)technologies or new consumer preferences. In CSM the cause-effect question is not

really valid, co-evolution is a combination of factors, only some are traceable to early origins and

butterfly effects. This co-evolution of agriculture and its matter-mind environment has over the past

century been characterised by large yield increases in Australia and the Netherlands, but also by a crisis

of thought about its future. Many of the processes can be explained by using various forms of system

thinking, some of which themselves hark back millennia. System thinking has oscillated between

reductionism and holism; it fruitfully divided issues of matter and mind only to recombine them again in

the course of time; it also separated observation and reason only to recombine them again. It accepted

and resisted uncertainty, eventually returning from a road emphasising control to one of participation.

Major changes in agriculture in the 20th century were made possible thanks to concepts from

reductionist and mechanistic thinking in HSM. However, it is increasingly acknowledged that „ parts‟

interact, that they learn, and that they cannot be seen as being independent from their context (CSM)

indefinitely. Past focus on statically defined systems needs to be balanced by a discussion on how

systems interact, and on how they are both host to smaller systems and component parts of larger ones:

how they combine matter (HSM) and mind (SSM) aspects into the more holistic embrace of CSM.

The curious tension of CSM is that uncertainty and different perceptions are likely to affect our world-

view for years to come. At the same time it is „ rather ‟ certain that agriculture is an interacting

combination of nature, soil and man in which mind issues are not separate from tangible matter.

Agriculture means different things to different stakeholders, but is definitely about more than just short-

term food supply. Farmers cannot ignore the functions of apparently useless fauna, nor can urban

populations ignore the fact that their own interests and lifestyles are intertwined with well-being andsustainability of agriculture. New processes and criteria are being developed to monitor and co-evolve

with system health; learning systems are required to reduce administrative rigidity and locked-in

teaching modes, so as to move from control to participation. Diversity in all its forms implies recurrence

of forms, it helps to better „ use‟ resources as well as to prepare for change in terms of mind and matter.

Most if not all successful development eventually leads to partially self-inflicted change. Examples of

lessons arising from the application of thermodynamic laws (and common sense) are those of ceteris

imparibus, lock-in and predator-prey cycles. They have more than just instrumental value: they have

significance for the world-view of man and for farming practices, agricultural research, policy-making

and teaching. Man is a part of nature rather than a ruler who controls it. History repeats itself, but never

in the same way, by combining processes of creation and destruction where mankind has choices to be

made, on priorities and allocation of resources to alternative coping strategies. This is perhaps themoral and central message of modern system thinking for farming beyond 2000.

Acknowledgements

Hans Lyklema, Piet Elenbaas, Wim Van Densen, Wiebe Aans, Wouter Joenje, Conny Almekinders,

Mario Giampietro, Brian Goodwin, Richard Bawden and Gerard de Zeeuw and many others have

contributed more to this chapter than they may be aware of themselves. Thanks are also due to Lenny

Boonen, Laura Cueva, Arthur Eveleens and Rinske Schiere who helped to get these ideas on paper

through a near infinite series of drafts.


23/27

23

References

Allen, T.F.H., Tainter, J.A. and Hoekstra, T.W. (1999). Supply-Side Sustainability, System Research

and Behavioral Science, 16: 402-427.

Ashby, R.W. (1958). General system theory as a new discipline. Society for General System Research,Gen. Syst, Yearb. 3, 1.

Auerbach, R. (1999). Design for participation in ecologically sound management of South Africa‟s

Mlazi River Catchment. PhD-thesis Wageningen, The Netherlands, 263 pp.

Brady, N.C. (1983). Agronomists' role in agriculture's revolution: Phase I. Agronomy Journal , 75: 401-

407.

Capra, F.(1997). The web of life, a new synthesis of mind and matter . Flamingo / Harper Collins,

London. 320 pp.

Chambers, R. Pacey, A. and Thrupp, L.A. (1989). Farmers first: farmer innovation and agricultural

research. London: Intermediate Technology Publications. 218 pp.

Checkland, P.B.,(1999). Soft Systems Methodology: a 30 year retrospective. John Wiley & Sons,

London. 330 pp.

Checkland, P.B. and Scholes P. (1990). Soft System Methodology in Action. Wiley. Chichester.

Cohen, J. and Stewart, I. (1994). The collapse of chaos:

4 systems thinkinh in agri

Documents